FUTURO on-device photo recognition

TinyCLIP ViT-8M / Text-3M
Source https://huggingface.co/onnx-community/TinyCLIP-ViT-8M-16-Text-3M-YFCC15M-ONNX
Pinned revision 9463a9c508a344c837ffefe9d724f3827bf2dc79
Base model https://huggingface.co/wkcn/TinyCLIP-ViT-8M-16-Text-3M-YFCC15M
MIT license: dist/models/tinyclip/LICENSE.txt
Modification: image encoder extracted from full quantized graph with ONNX 1.17.0; trained weights unchanged. Text embeddings precomputed from 99 label prompts (86 foods, 13 non-food), with Transformers.js 3.8.1 (Apache-2.0), which is used for development only and not shipped in the browser.

ONNX Runtime Web 1.21.0, WASM-only entrypoint (Microsoft, MIT): dist/vendor/ort/LICENSE.txt
All model and inference assets are served from this site. No photo data leaves the browser. Models are not food-safety, allergy, quantity, or nutrition detectors. Candidates must be confirmed manually.